本文是一份可复现的实测教程，用固定随机种子从 LibriSpeech test-clean 和 test-other 采样，在 Apple M4 的 CPU 上对比 Cactus Compute 的 16.9 MB 语音识别模型 Whistle 与 Whisper tiny、base。作者采用配对说话人聚类 bootstrap 区间并做 Holm 校正后发现：在 test-other 上 Whistle 显著优于 Whisper base 的贪心解码，但与 beam search 解码相比没有显著差距；速度优势存在，但音频越长差距越小。教程还测试了 30 秒单次窗口截断问题、流式延迟、时间戳、关键词偏置、多语言、噪声、嵌入以及语音调用工具。

AI 服务暂时不可用，以下为来源摘要，待恢复后补全翻译：arXiv:2610.28901v1 Announce Type: new Abstract: We consider cooperative and iterative tasks involving multiple agents whose interactions are modeled as a dynamic game with shared constraints. We propose a general learning framework applicable to nonlinear and nonconvex dynamic games that uses previously executed feasible trajectories to construct a sampled joint safe set and player-specific cost-to-go functions. At each time step, the agents jointly select a previously sampled terminal state and a Generalized Nash equilibrium used to reach it. A new solution is adopted only if it satisfies a cooperative acceptance rule agreed upon by the players. The framework accommodates individual non-degradation requirements and selection criteria such as weighted costs and…

该论文提出 FLEX（灵活特征编辑流水线），利用语言模型根据人类的自然语言纠正，迭代式地精炼机器人对任务特征的内部表征。与以往只允许“添加”特征的编辑方法不同，FLEX 通过添加缺失特征、泛化过于具体的特征、屏蔽无关特征三种互补操作，避免特征集过度指定或范围过窄。作者在二维导航、7 自由度 Franka 桌面操作任务以及真实机器人实验中对 FLEX 进行了评估，结果显示三种操作在交互场景中都不可或缺，且 FLEX 在保持表征紧凑、可迁移的同时，取得了更高的归一化奖励。该工作已被 CoRL 2026 接收。

该论文提出一种参数空间记忆框架，将每段机器人操作片段存储为固定大小的神经网络权重增量，而非随视频帧数增长的显式数据。权重增量经学习到的投影器转换为 token 嵌入，用于驱动语言模型问答与关节指令回归，在 RoboMME 基准上初步验证了有效性。

该论文提出Real2Sem2Real（R2Sem2R）流水线，仅凭一段单目RGB人类演示视频，就能为同一类别的多个语义对象实例生成数以千计的机器人演示数据。系统用DINOv3蒸馏的高斯泼溅跟踪物体SE(3)位姿，并将由少量照片生成的“数字表亲”与演示物体做语义对齐，再把手部动作映射为可行抓取，最后在IsaacSim或MuJoCo中批量生成数据。在两台真实机器人和四项任务上，用该方法训练的策略在多样对象实例上达到或超过遥操作数据训练的策略。

该论文提出 T-WLA 世界-语言-动作架构，通过共享视觉记忆同时驱动基础动作生成与轻量级触觉修正，使机器人在视觉策略更新之前就能对接触变化做出快速调整。在指关节运动预测任务中，该方法将 MPJPE 降低 30%，并在四项真实机器人灵巧操作任务中取得约 50.0% 的平均成功率。

AI 服务暂时不可用，以下为来源摘要，待恢复后补全翻译：arXiv:2610.28791v1 Announce Type: new Abstract: Target-based extrinsic calibration estimates the rigid transform between sensors from a few observations of a known target, and its result is commonly assessed by the root-mean-square error of the fitted residuals. However, this scalar error is computed on the same observations and therefore cannot indicate whether to accept the result, re-extract, recollect or add observations. This paper presents VeriCalib, which formulates extrinsic calibration as a verifiable estimation problem and reports a calibration state. The key insight is a two-level extrinsic compensation mechanism, in which the extrinsic compensates the rigid part of each observation error at the residual level, and rotation and translation compensate…

该论文提出一种用于物理人机交互（pHRI）的新框架，将鲁棒力/力矩（wrench）观测器与位姿-力混合控制器结合，通过神经网络补偿摩擦与未建模动态、利用任务-零空间分解抑制人类扰动，从而在不使用力/力矩传感器的情况下实现位姿与力的可靠跟踪。

Nuo Chen 与 Ya-Jun Pan 提出一个元学习框架，用于人机交互中的交互模式识别与自适应控制。该框架用元学习训练的神经网络替代人工设定的切换阈值来判断人类意图，被动模式识别率由 90.1% 提升至 99.1%，主动模式由 79.8% 提升至 97.1%；同时引入自适应元学习控制器补偿人类扰动，并用多轨迹元学习算法联合学习环境基表示与力相关参数，以应对未建模动力学。

AI 服务暂时不可用，以下为来源摘要，待恢复后补全翻译：arXiv:2610.28712v1 Announce Type: new Abstract: First-person humanoid teleoperation ties access to side and rear targets to operator rotation. We present coupled visual--motor remapping, pairing a panoramic azimuth warp with its inverse upper-limb yaw transform. A working prototype implements Global Zoom (GZ) and Focus+Context (F+C). In a within-subject study (N=8); 72 trials), GZ had the largest completion count (23/24; Fixed: 20/24; F+C: 21/24), whereas F+C was preferred by 6/8 participants. Therefore, completion and preference favored different mappings in this sample.

AI 服务暂时不可用，以下为来源摘要，待恢复后补全翻译：arXiv:2610.28701v1 Announce Type: new Abstract: Accurate ground plane normal estimation is essential for odometry in spherical mobile mapping systems (SMMS), which face challenges from rapid rotations and limited ground coverage. We propose two RANSAC-based approaches: single-scan estimation with normal vector scoring and multi-step validation, and LiDAR-inertial odometry (LIO)-based estimation using a global 3D map. Evaluated across diverse real-world datasets against terrestrial laser scanner ground truth, results show that LIO-based estimation significantly outperforms single-scan methods in accuracy and robustness. Integrating ground normal estimates into the motion model improves pose estimation, supporting autonomous navigation in challenging and hazardou…

该论文提出 DAFIC 框架，将单目深度信息引入小样本图像分类，通过 RGB-D 自监督预训练、深度引导几何自适应模块与空间双向块匹配，缓解背景杂乱和 RGB 外观相似造成的视觉歧义。该工作已被 WACV 2027 接收，并在四个基准上取得优于现有方法的表现。

AI 服务暂时不可用，以下为来源摘要，待恢复后补全翻译：arXiv:2610.28843v1 Announce Type: new Abstract: Visual question answering requires models to assign attributes and relations to the correct objects, a problem known as visual binding. Structured prompts make these assignments explicit through labeled fields. Yet an improvement from the full prompt does not reveal which labels help or how they interact. First, we isolate field-label effects across ten models and six tasks by holding the image and claim fixed. These comparisons reveal differences that overall accuracy can hide. With a relation label present, adding object-role labels helps Qwen3 but hurts Qwen2 on Spatial. Second, we compare auxiliary hints that preserve or swap the question's role assignments against a no-hint baseline. Across two image sets, Qw…

AI 服务暂时不可用，以下为来源摘要，待恢复后补全翻译：arXiv:2610.28835v1 Announce Type: new Abstract: Conventional visual sensing faces a fundamental dilemma: information that should remain private today may need to be recovered tomorrow. This challenge is becoming increasingly important as cameras become pervasive in our daily surroundings. Existing safeguards either protect visual content only after an image is formed or suppress information during capture in ways that prevent its later recovery. Here, we show through an information-theoretic analysis that privacy and recoverability cannot be achieved simultaneously in static optical systems. Information preserved in a measurement for later reconstruction also remains available to an observer of that measurement, limiting the privacy such systems can provide. To…

AI 服务暂时不可用，以下为来源摘要，待恢复后补全翻译：arXiv:2610.28829v1 Announce Type: new Abstract: Instruction-guided video generative models offer a scalable solution for stress-testing autonomous driving systems by simulating diverse environmental conditions. However, driving scenes operate within a dense Operational Design Domain (ODD) governed by strict traffic semantics. When applied to global driving-condition edits, general-purpose instruction-guided video editors frequently corrupt these semantics, leading to structural hallucinations and ``temporal shock.'' This work distills and protects traffic semantics during video editing at three levels. First, we construct a hybrid dataset by combining temporally aligned real-world videos with synthetic edits, balancing foundational scene structure with unambigu…

AI 服务暂时不可用，以下为来源摘要，待恢复后补全翻译：arXiv:2610.28827v1 Announce Type: new Abstract: We present WorldStream, an open-ended 3D world model with three simultaneous capabilities: it extends playable terrain beyond any pre-specified map, acts on open-ended language in the running world, and restores visited places with their edits. A hierarchical planner over a typed decision model emits structured decision state: Macro fixes settlement, ecology, transport, and hydrological structure across a group of tiles; Micro specializes that contract into tile semantics under measured terrain and accepted neighbor commitments. Deterministic local code compiles accepted plans into terrain, roads, settlements, vegetation, and entities. Streaming separates their reach: plans run ahead along the predicted corridor,…

arXiv 上一篇被 NeurIPS 2026 接收的论文提出 Register Anything Model（RAM），这是一种在室内、室外和物体等多领域数据上联合训练的统一点云配准模型，可跨传感器和场景鲁棒估计 6-DoF 变换。其核心是几何感知归一化策略，将点云自适应体素化并归一化到标准几何空间，再配合由粗到细的分层框架和高效的几何 Transformer。在 12 个基准上，该方法在内点率和配准召回率上达到最佳水平，甚至优于针对每个单独领域训练的模型。

arXiv 论文 SAIL 提出一个可微框架，将 Moran's I、Geary's C 等经典空间自相关统计量推广为图上的可学习算子：它联合学习特征投影与节点间邻域亲和度，输出可解释的局部与全局空间一致性描述子，既能用于监督预测，也能无监督地发现空间自相关的模式与热点。

本文提出 JUICE，一个在冻结的图像与文本预训练表征空间上运行的单一流匹配 Transformer，两个模态各有独立噪声水平，通过哪些模态保持干净、哪些需要生成的设定来切换图像生成、图像描述、视觉问答与联合生成等任务。在 2.1B 参数、224x224 分辨率且不使用语言模型初始化的条件下，它最高达到 GenEval 0.88 与 VQAv2 73.1，并在需要双模态互相配合的迷宫路径规划任务中显著优于独立采样。

本文提出SurgReasonSeg，一种复杂度感知的手术视频推理分割方法，利用数字孪生表示、双LoRA适配器和强化学习路由，在实时性下显著提升语义、空间和时间查询的分割性能，并引入新基准EndoVis-ReasonSeg。

一篇被 WACV 2027 接收的论文对 MS1MV3 人脸识别训练集进行受控审计，识别出五类残留数据缺陷，并通过基于 IResNet-50 与 ArcFace 的受控训练验证：针对性地修正缺陷可在减少优化步数的同时维持甚至提升识别性能，其中重复图像去除的收益最为明确。

AI 服务暂时不可用，以下为来源摘要，待恢复后补全翻译：arXiv:2610.28840v1 Announce Type: new Abstract: Early-grade reading screening is critical for identifying children at risk of reading difficulty. Existing AI-based screeners are often limited in the scope of their analyses and evaluation metrics, or are proprietary systems that lack transparent model specifications and evaluation. We present KIRA, a dual-pathway screening system that pairs word-level fluency scoring with detailed phoneme-level error detection. We adapt pretrained word and phoneme recognizers to children's speech, reducing word error rate from 13.88% to 8.84% and phoneme error rate from 41.66% to 23.00%. Powered by these child-adapted models, we implement Word Reading (WRE) and Oral Reading Fluency (ORF), two foundational tasks in K-2 screening.…

一篇新论文研究如何在多语言预训练中显式利用平行数据的结构，发现单独使用平行预训练（PPT）并不总能提升语言理解能力，但采用“先用PPT构建跨语言基础、再用标准预训练进行适配”的两阶段方法，在高资源和低资源场景下都能稳定超过纯标准预训练。

研究人员发布 MADARTB，这是首个覆盖贝鲁特、开罗、多哈、拉巴特和突尼斯五种城市方言的大规模平行阿拉伯语依存树库，并包含现代标准阿拉伯语（MSA）对照版本。该资源共包含 5,000 个采用 CATiB 依存标注体系标注的句子，用于比较各方言与 MSA 之间的句法差异，以及监督式解析器与大型语言模型在跨方言依存解析上的表现。

来自 arXiv 的一篇新论文提出了 Lift-Bench 与 Lift-Train，用于评估和训练语言模型的指令跟随能力。其核心创新是把约束限定在文档的特定章节内，从而更精细地控制约束密度；基准同时包含可程序验证的约束和由评审判断的约束，并同时提供主任务约束和指令跟随约束，以缓解浅层奖励黑客问题。作者发现前沿模型在较难的子集上表现吃力，而在 Lift-Train 上训练比在常见的 IFTrain 上训练更不易出现奖励黑客。

该论文研究大型语言模型在处理法律叙事时，其内部激活中是否存在与CATO全部26个法律要素相对应的表征。作者在Qwen3-1.7B、Qwen3-8B和Gemma-2 2B三个开放模型上进行线性可解码性与稀疏自编码器稀疏定位实验，发现全部26个要素在三个模型中均可线性解码，最佳层测试AUC为0.86至1.00，并报告了78个模型特定的层-特征坐标。

该论文指出现有的LLM定价智能体合谋评估只关注结果，却忽略了合谋的关键——对作弊者的惩罚执行。作者提出一种审计方法，通过在对市场的反事实副本中强制某一成员作弊来检验执行能力，并研究强化学习在定价与交易中究竟教会了智能体什么。

AI 服务暂时不可用，以下为来源摘要，待恢复后补全翻译：arXiv:2610.28772v1 Announce Type: new Abstract: Psychotherapy fidelity assessment measures how faithfully a therapist adheres to a treatment protocol, a critical determinant of positive patient outcomes. However, fidelity assessment is bottlenecked by manual expert review, rendering the process labor-intensive and unscalable. To scale therapy supervision and support clinician self-learning, we present an integrated application of large language models (LLMs) that automates psychotherapy fidelity workflows and generates actionable supervisory feedback from authentic, long-form clinical transcripts. We evaluate three open-weight models across multiple task decomposition configurations and reasoning strategies, demonstrating that evidence-first reasoning improves…

研究者提出 SkillEval 评估框架，利用大语言模型的元认知能力从题目中自动识别可解释的技能，并结合心理测量学中的认知诊断模型，为每个模型估算各项技能的掌握水平，从而替代传统上把所有题目聚合成单一分数的粗粒度评测方式。

Joss Moorkens 与 Monja Jannet 撰写的一篇书章，为翻译领域引入数据伦理议题，讨论版权、数据所有权与偏见等普遍性问题，以及少数语言社群和数字鸿沟两侧群体所面临的特殊伦理挑战，并探讨责任、监管以及翻译研究界对数据伦理日益增长的关注。